So, how can you have successful Business Intelligence implementations?
In my last post “Why BI projects fail?” we looked at some of the fundamental differences between core IT system and BI system implementations. In this post, I will outline proven approaches to successfully implement BI solutions and show how that is different from implementing core IT systems.
To keep things simple, we will analyse implementations in three phases. And we will see that there are significant differences between BI and core IT implementations in all three phases.
The initial choices you make will largely impact your BI implimentation
The fundamental difference in the pre-implementation phase is the choice of the main stakeholders or internal champions of the implementation.
Core IT system implementations are typically led by the CIO and sponsored by a business unit leader. This makes sense because core IT implementations could require new hardware and software and they usher in a new way of working for most employees.
BI implementations, on the other hand, should be led by a business unit leader (ideally a C-suite executive) and be supported by IT because the utility and value of a BI system is ultimately realised by executives. IT should play a role in selecting the product, particularly with regards to issues involving interfaces, security and configuration.
The choice of the stakeholders driving a BI project is perhaps the most important factor in determining its success. The right person will clear obstacles, provide direction and drive results.
2Implementation process and focus
BI implementations should be agile and based on rapid iterations in contrary to core implementations
Core IT system implementations are generally long and require a large implementation team from the vendor and a cross-functional team from the business. Project plans, budgets, and timelines are drawn up, and agreed upon with a Steering Committee… and you know what happens then…. Core systems are complex in nature and need to be configured to implement organizational processes. Therefore, by definition, you cannot “partially” implement a core system. The focus should be on ensuring the end-to-end process is well mapped out and information flows seamlessly across all aspects of the process. If not, you will see manual workarounds creep in.
BI implementations, on the other hand, should be agile and based on rapid iterations which quickly deliver specific insights to management. The keyword there is quick – not holistic. Whilst this may sound counter-intuitive it is a very effective way of delivering immediate value to the business, because more often than not, you can measure performance at a sub-process level – such as conversions in sales or lead generation in marketing – while the rest of the process is being built. (NOTE – DO NOT aim for a full blown implementation that covers all aspects of your analytical needs. DO measure what’s important because what gets measured gets improved and this feeds into a virtuous cycle as the rest of the process is being built.)
Usually, a good starting point is to look at existing reports that are being used by management. The initial aim of a BI project should be to ensure that those reports are generated as quickly and efficiently as possible. The system should encourage stakeholders to move from consuming information from a “periodic” basis to an “on-demand” basis. This will ensure that management has access to current information when they make decisions. Once the initial system has been set up, management will have the ability to review and revise what’s being measured, how often and how effectively, and this leads to the next iteration.
For BI solutions, post implementation enables the management to harness the real benefit out of it
Once core IT systems have been implemented, training and change management are the key areas of focus for the organisation. Training programs are put in place to ensure employees are conversant with the new system and any new processes that are warranted by the implementation are in place. Change management projects are set up to ensure that the staff is not resisting the switch over. These are necessary because it is generally very hard to make significant changes to the system post implementation due to the extensive involvement of humans, machines and processes.
For BI solutions, the post-implementation focus should be on the impact it has on the ability of management to make better decisions. The key question should be “Are we measuring the right metrics which enable us to make better decisions?”. The tool by and of itself should be “invisible” and require minimum training. It should seamlessly blend into the organisational fabric. By definition, the solution should promote agility and provide a platform for the executives to constantly tweak their metrics and framework.
At this point, you may be wondering about data validation, data cleansing, and business model mapping. Though these factors are important to a BI implementation and form the operational layer of an implementation, the focus of this post is on getting the big picture right.
In summary, the most successful BI implementations are the ones that have the three key elements:
- They are driven by a business executive
- They utilize an Agile implementation process with short-burst iterations
- They focus on validating metrics and their relevance to better decision making
In my final post for this series, I will discuss the operational aspects of implementing a BI solution.